Chemodiversity of Brevetoxins and Other Potentially Toxic Metabolites Produced by Karenia spp. and Their Metabolic Products in Marine Organisms

In recent decades, more than 130 potentially toxic metabolites originating from dinoflagellate species belonging to the genus Karenia or metabolized by marine organisms have been described. These metabolites include the well-known and large group of brevetoxins (BTXs), responsible for foodborne neurotoxic shellfish poisoning (NSP) and airborne respiratory symptoms in humans. Karenia spp. also produce brevenal, brevisamide and metabolites belonging to the hemi-brevetoxin, brevisin, tamulamide, gymnocin, gymnodimine, brevisulcenal and brevisulcatic acid groups. In this review, we summarize the available knowledge in the literature since 1977 on these various identified metabolites, whether they are produced directly by the producer organisms or biotransformed in marine organisms. Their structures and physicochemical properties are presented and discussed. Among future avenues of research, we highlight the need for more toxin occurrence data with analytical techniques, which can specifically determine the analogs present in samples. New metabolites have yet to be fully described, especially the groups of metabolites discovered in the last two decades (e.g tamulamides). Lastly, this work clarifies the different nomenclatures used in the literature and should help to harmonize practices in the future.


Introduction
The genus Karenia belongs to the clade of the alveolates (with the ciliates and the apicomplexa). It includes 12 known marine species from oceanic and coastal areas. These eukaryotes were formerly classified as microalgae, and belong to the group of photosynthetic protists (autotrophs) [1]. Among the genus Karenia, some species of this dinoflagellate can proliferate and form dense blooms that cause multiple dysfunctions of the hydrosystem, but also have a strong toxic potential for organisms in the food web [2,3]. These massive blooms, referred as "red tides", can be visible and color the water. Brevetoxins (BTXs), which are produced by Karenia brevis (C.C. Davis) Gert Hansen and Moestrup, 2000, belong to the main group of marine biotoxins associated with these events. These neurotoxins can cause significant mortality in fish, sea birds, and marine mammals [4][5][6]. Humans can also be exposed to BTX through food, leading to human of numerous new metabolites. For many of them, chemical structure elucidation relied on a stringent study of their mass spectrometry fragmentation profiles. Confirmation of these structures with more thorough elucidation tools, such as nuclear magnetic resonance (NMR), would require large quantities of samples for isolation and purification purposes, making this goal difficult to achieve.
The aim of this review is to summarize the knowledge collected in recent decades about the potentially toxic metabolites produced by Karenia spp. or reported in other marine organisms. Therefore, metabolites obtained by chemical synthesis (partially or fully from existing compounds) or from in vitro metabolism studies will not be consi dered [51][52][53][54][55]. We will present and discuss the structures and physicochemical properties of the metabolites reported in the literature. The accumulation and biotransformation of these potentially toxic metabolites by marine organisms will be discussed in depth and will help to identify and highlight future avenues of research. As previously mentioned, the use of different nomenclatures can be disruptive and it is not always easy to switch between the different names used, especially for BTXs. We will tackle this aspect by reporting all the names used in the literature and by preferentially using those which seem to achieve a better overall consistency.

Metabolites Produced by Karenia spp.
Among the different Karenia species reported, K. brevis is known to produce BTXs, brevisin, brevisamide, brevenals and tamulamides, whereas brevisulcenals and brevisulcatic acids can be produced by Karenia brevisulcata (F.H. Chang, 1999) [24,[56][57][58], whereas gymnodimine-B and -C were first isolated from Karenia selliformis A.J. Haywood, K.A. Steidinger & L. MacKenzie, 2004 [59]. All these groups of metabolites possess different structure skeletons that give them their physicochemical properties (Table 1). To allow for a better follow-up of the numerous compounds reported in this review, an identification number (ID) was attributed to each of them in the following parts of this article (text, tables and figures).   Figure 1. Geographical distribution of potentially toxic metabolites individually and formally identified (blue color for country), displayed by type of samples collected (forms) and by group to which they belong (colors of forms). Map generated using R software and the ggplot2 package [83,84]. Figure 1. Geographical distribution of potentially toxic metabolites individually and formally identified (blue color for country), displayed by type of samples collected (forms) and by group to which they belong (colors of forms). Map generated using R software and the ggplot2 package [83,84].

Brevetoxins
BTXs are lipid-soluble and thermostable polycyclic polyethers produced by the dinoflagellate Karenia brevis [13]. In laboratory conditions, the production of BTX-2 (ID #8) has also been shown for Karenia papilionacea A.J. Haywood and K.A. Steidinger, 2004 [85]. These toxins are polyethers with ladder-like structures similar to ciguatoxins (CTXs) (produced by the dinoflagellate Gamberdiscus). All polyether compounds produced by K. brevis are thought to derive from polyepoxide precursors [86,87]. A wide range of BTXs have been identified from K. brevis cultures (Table 1). NMR and X-ray crystallography were the analytical techniques implemented to establish their structures. A-type and B-type BTXs (BTX-A and BTX-B) are the two base skeletons identified ( Figure 2). BTX-1 (ID #1) and BTX-2 are the parent algal toxins of the BTX-A and BTX-B types, respectively [88]. The molecular formula of BTX-1 is C 49 H 70 O 13 and its structure is composed of 10 rings [60,66]. The molecular formula of BTX-2 is C 50 H 70 O 14 and its structure is composed of 11 rings [14,64]. Both structures carry a lactone function that is thought to be responsible for their biological activity [88]. BTX-2 is the major toxin isolated from K. brevis and the one produced by the largest number of species of the genus Karenia [88].
BTXs are neurotoxins that primarily target the voltage-gated Na+ channels (Na V ) [11]. The neurological nature of the primarily symptoms observed in humans and animals can be explained by the specific affinity of BTXs for Na V channel subtypes and the tissue distribution of Na V channels. The symptoms also involve gastrointestinal and cardiovascular systems. Data on acute toxicity in animals are very limited. The only toxicity study by oral administration allows for an estimation of the median lethal doses (LD 50 ) for BTX-2 and BTX-3 in female mice [14]. BTX-3 was about 10 times more toxic than BTX-2 by oral administration, with LD 50 of 520 and 6600 µg/kg bw, respectively. However, by intraperitoneal injection, the two toxins were almost equipotent.

Hemibrevetoxins
Hemibrevetoxins (Hemi-BTX-A, -B, -C with respective ID of #17, #18, #19), cyclic ethers biosynthesized by K. brevis, were first isolated by Shimizu [68] in 1982 and were reported to have weak ichthyotoxicity. Later, Prasad and Shimizu elucidated the chemical structure of Hemi-BTX-B [69] (Figure 3). Its molecular formula is C 28 H 42 O 7 , and is about half that of BTXs, hence its name ( Table 1). The authors suggested that hemibrevetoxins could play an important role in the biosynthesis of BTXs. They also noted cytotoxicity at a concentration of 5 µmol on cultured mouse neuroblastoma cells, and a characteristic rounding of cells, as for BTX-A and BTX-B. Hemi-BTX-A and -C were isolated; however, their chemical structures seem to have never been fully elucidated [68,69].

Brevenals
In 2004, brevenal (ID #20) and its dimethyl acetal (ID #21) were first isolated and characterized from K. brevis cells collected in Florida, USA ( Figure 4). These fat-soluble compounds, with molecular formulae C 39 H 60 O 8 and C 41 H 66 O 9 , respectively, are composed of five ether rings and possess an aldehyde function [16,70,81]. Brevenals are not toxic for fish and antagonize the effects of BTXs. Both competitively displace BTX from its binding site in rat brain synaptosomes. Moreover, brevenal inhibits the bronchoconstriction induced by BTX-2 (ID #8) and BTX-3 (ID #9) in a sheep assay [90]. In 2006, the first synthesis of brevenal was accomplished [92]. This work led to a revision of the structure of brevenal, which corresponds to the C26 epimer of the initially proposed natural product structure (Figure 4)

Brevisamide
In 2008, brevisamide (ID #22) was isolated from K. brevis [71]. Its molecular formula is C 18 H 29 NO 4 and its structure is composed of a monocyclic ether possessing an amide function. It is also composed of a tetrahydropyran ring bearing a 3,4-dimethylhepta-2,4 dienal side chain ( Figure 5). To date, no other analog has been discovered. Brevisamide was synthetized by several authors using different strategies [93][94][95]. Brevisamide may be the biosynthetic precursor of brevisin, based on their respective structures [95]. The toxicity of this metabolite remains unknown, due to the absence of any studies.   The structural elucidation of BTX-1 (ID #1) and BTX-2 (ID #8) facilitated the resolution of the structure of all subsequently discovered all related analogs ( Figure 2). Each family consists of several compounds, differentiated by the "J-ring" in BTX-A (BTX-1, BTX-7, BTX-10 with respective ID of #1, #2, #3) and at the "K-ring" in BTX-B (BTX-2, BTX-3, BTX-5, BTX-6, BTX-9 reported to have weak ichthyotoxicity. Later, Prasad and Shimizu elucidated the chemical structure of Hemi-BTX-B [69] (Figure 3). Its molecular formula is C28H42O7, and is about half that of BTXs, hence its name ( Table 1). The authors suggested that hemibrevetoxins could play an important role in the biosynthesis of BTXs. They also noted cytotoxicity at a concentration of 5 µmol on cultured mouse neuroblastoma cells, and a characteristic rounding of cells, as for BTX-A and BTX-B. Hemi-BTX-A and -C were isolated; however, their chemical structures seem to have never been fully elucidated [68,69] Figure 3. Chemical structure of hemibrevetoxin-B (ID #18).

Brevenals
In 2004, brevenal (ID #20) and its dimethyl acetal (ID #21) were first isolated and characterized from K. brevis cells collected in Florida, USA ( Figure 4). These fat-soluble compounds, with molecular formulae C39H60O8 and C41H66O9, respectively, are composed of five ether rings and possess an aldehyde function [16,70,81]. Brevenals are not toxic for fish and antagonize the effects of BTXs. Both competitively displace BTX from its binding site in rat brain synaptosomes. Moreover, brevenal inhibits the bronchoconstriction induced by BTX-2 (ID #8) and BTX-3 (ID #9) in a sheep assay [90]. In 2006, the first synthesis of brevenal was accomplished [92]. This work led to a revision of the structure of brevenal, which corresponds to the C26 epimer of the initially proposed natural product structure ( Figure 4).

Brevisamide
In 2008, brevisamide (ID #22) was isolated from K. brevis [71]. Its molecular formula is C18H29NO4 and its structure is composed of a monocyclic ether possessing an amide function. It is also composed of a tetrahydropyran ring bearing a 3,4-dimethylhepta-2,4 dienal side chain ( Figure 5). To date, no other analog has been discovered. Brevisamide was synthetized by several authors using different strategies [93][94][95]. Brevisamide may be the biosynthetic precursor of brevisin, based on their respective structures [95]. The toxicity of this metabolite remains unknown, due to the absence of any studies.

Brevisin
In 2009, brevisin (ID #23) was identified from K. brevis cells [72,96]. Its structure consists of two separate fused polyether ring assemblies linked by a methylene group. One of the polycyclic moieties was conjugated with an aldehyde side chain ( Figure 6). Brevisin inhibits the binding of [ 3 H]-BTX-3 on sodium channels in rat brain synaptosomes. In 2011, the total synthesis of brevisin was accomplished [97]. To date, studies regarding this secondary metabolite are limited and further investigation is required.

Tamulamides
Tamulamides, cyclic polyethers with a ladder-like chemical structure, were isolated from K. brevis cultures [73]. Both are composed of seven cyclic ethers with an amide and aldehyde side chain (Figure 7). The molecular formulae of Tamulamide-A (Tam-A; ID #24) site in rat brain synaptosomes. Moreover, brevenal inhibits the bronchoconstriction induced by BTX-2 (ID #8) and BTX-3 (ID #9) in a sheep assay [90]. In 2006, the first synthesis of brevenal was accomplished [92]. This work led to a revision of the structure of brevenal, which corresponds to the C26 epimer of the initially proposed natural product structure ( Figure 4).  Figure 4. Chemical structures of brevenal (ID #20) and dimethyl acetal brevenal (ID #21).

Brevisamide
In 2008, brevisamide (ID #22) was isolated from K. brevis [71]. Its molecular formula is C18H29NO4 and its structure is composed of a monocyclic ether possessing an amide function. It is also composed of a tetrahydropyran ring bearing a 3,4-dimethylhepta-2,4 dienal side chain ( Figure 5). To date, no other analog has been discovered. Brevisamide was synthetized by several authors using different strategies [93][94][95]. Brevisamide may be the biosynthetic precursor of brevisin, based on their respective structures [95]. The toxicity of this metabolite remains unknown, due to the absence of any studies.

Brevisin
In 2009, brevisin (ID #23) was identified from K. brevis cells [72,96]. Its structure consists of two separate fused polyether ring assemblies linked by a methylene group. One of the polycyclic moieties was conjugated with an aldehyde side chain ( Figure 6). Brevisin inhibits the binding of [ 3 H]-BTX-3 on sodium channels in rat brain synaptosomes. In 2011, the total synthesis of brevisin was accomplished [97]. To date, studies regarding this secondary metabolite are limited and further investigation is required.

Tamulamides
Tamulamides, cyclic polyethers with a ladder-like chemical structure, were isolated from K. brevis cultures [73]. Both are composed of seven cyclic ethers with an amide and aldehyde side chain ( Figure 7). The molecular formulae of Tamulamide-A (Tam-A; ID #24)

Brevisin
In 2009, brevisin (ID #23) was identified from K. brevis cells [72,96]. Its structure consists of two separate fused polyether ring assemblies linked by a methylene group. One of the polycyclic moieties was conjugated with an aldehyde side chain ( Figure 6). Brevisin inhibits the binding of [ 3 H]-BTX-3 on sodium channels in rat brain synaptosomes. In 2011, the total synthesis of brevisin was accomplished [97]. To date, studies regarding this secondary metabolite are limited and further investigation is required. site in rat brain synaptosomes. Moreover, brevenal inhibits the bronchoconstriction induced by BTX-2 (ID #8) and BTX-3 (ID #9) in a sheep assay [90]. In 2006, the first synthesis of brevenal was accomplished [92]. This work led to a revision of the structure of brevenal, which corresponds to the C26 epimer of the initially proposed natural product structure ( Figure 4).  Figure 4. Chemical structures of brevenal (ID #20) and dimethyl acetal brevenal (ID #21).

Brevisamide
In 2008, brevisamide (ID #22) was isolated from K. brevis [71]. Its molecular formula is C18H29NO4 and its structure is composed of a monocyclic ether possessing an amide function. It is also composed of a tetrahydropyran ring bearing a 3,4-dimethylhepta-2,4 dienal side chain ( Figure 5). To date, no other analog has been discovered. Brevisamide was synthetized by several authors using different strategies [93][94][95]. Brevisamide may be the biosynthetic precursor of brevisin, based on their respective structures [95]. The toxicity of this metabolite remains unknown, due to the absence of any studies. Figure 5. Chemical structure of brevisamide (ID #22).

Brevisin
In 2009, brevisin (ID #23) was identified from K. brevis cells [72,96]. Its structure consists of two separate fused polyether ring assemblies linked by a methylene group. One of the polycyclic moieties was conjugated with an aldehyde side chain ( Figure 6). Brevisin inhibits the binding of [ 3 H]-BTX-3 on sodium channels in rat brain synaptosomes. In 2011, the total synthesis of brevisin was accomplished [97]. To date, studies regarding this secondary metabolite are limited and further investigation is required.

Tamulamides
Tamulamides, cyclic polyethers with a ladder-like chemical structure, were isolated from K. brevis cultures [73]. Both are composed of seven cyclic ethers with an amide and aldehyde side chain ( Figure 7). The molecular formulae of Tamulamide-A (Tam-A; ID #24)

Tamulamides
Tamulamides, cyclic polyethers with a ladder-like chemical structure, were isolated from K. brevis cultures [73]. Both are composed of seven cyclic ethers with an amide and aldehyde side chain ( Figure 7). The molecular formulae of Tamulamide-A (Tam-A; ID #24) and Tamulamide-B (Tam-B; ID #25) are C 35 H 45 NO 10 and C 34 H 43 NO 10 , respectively (Table 1). Tam-A and Tam-B compete with [ 3 H]-BTX-3 for its binding site on rat brain synaptosomes. Neither showed toxicity to fish at doses up to 200 nM and cause only slight bronchoconstriction in sheep pulmonary assays [73].  Tamulamide

Gymnocins
Gymnocins are polyethers biosynthesized by K. mikimotoi. Gymnocin-A (ID #26) was the first analog discovered from Japanese cultures of this dinoflagellate, isolated at Kushimoto Bay [56] (Figure 1). Structure elucidation revealed 14 saturated ether rings, bearing a 2-methyl-2-butenal side chain ( Figure 8). In 2003, this analog was fully synthetized by Tsukano and Sasaki [98]. Gymnocin-A carboxylic acid (ID #27), Gymnocin-A2 (ID #28), Gymnocin-B (ID #29; composed of 15 adjacent rings ending with the same 2-methyl-2-butenal side chain as gymnocin-A) were isolated and elucidated from K. mikimotoi [57,58] (Table 1). Conventional fish assays showed that Gymnocin-A and Gymnocin-B are weakly toxic compared to BTX-B [56,57]. However, the authors highlighted that fish were directly exposed to dissolved gymnocins, whereas during red tide events, K. mikimotoi cells stuffed the fish gills, allowing direct contact of the gymnocins with the gills. The extremely low solubility of gymnocins to water could prevent them from reaching the fish gills during assays. This mechanism could be involved in the fish mortalities observed in the field, whereas conventional fish assays do not reveal significant toxicity. Gymnocin-A carboxylic acid and Gymnocin-A2 showed moderate cytotoxicity against P388 cells [58].

Gymnocins
Gymnocins are polyethers biosynthesized by K. mikimotoi. Gymnocin-A (ID #26) was the first analog discovered from Japanese cultures of this dinoflagellate, isolated at Kushimoto Bay [56] (Figure 1). Structure elucidation revealed 14 saturated ether rings, bearing a 2-methyl-2-butenal side chain ( Figure 8). In 2003, this analog was fully synthetized by Tsukano and Sasaki [98]. Gymnocin-A carboxylic acid (ID #27), Gymnocin-A2 (ID #28), Gymnocin-B (ID #29; composed of 15 adjacent rings ending with the same 2-methyl-2butenal side chain as gymnocin-A) were isolated and elucidated from K. mikimotoi [57,58] ( Table 1). Conventional fish assays showed that Gymnocin-A and Gymnocin-B are weakly toxic compared to BTX-B [56,57]. However, the authors highlighted that fish were directly exposed to dissolved gymnocins, whereas during red tide events, K. mikimotoi cells stuffed the fish gills, allowing direct contact of the gymnocins with the gills. The extremely low solubility of gymnocins to water could prevent them from reaching the fish gills during assays. This mechanism could be involved in the fish mortalities observed in the field, whereas conventional fish assays do not reveal significant toxicity. Gymnocin-A carboxylic acid and Gymnocin-A2 showed moderate cytotoxicity against P388 cells [58].

Gymnodimines (GYMs)
Initially, gymnodimine-A (GYM-A; ID #30) was isolated from extracts of New Zealand oysters (Tiostrea chilensis) [24] (Figure 1). Isolation of the active substance made it possible to determine its molecular formula as C32H45NO4 (Table 1), whereas its absolute stereochemistry was later determined by X-ray crystallography (Figure 9) [99]. The origin of this toxin was first

Gymnodimines (GYMs)
Initially, gymnodimine-A (GYM-A; ID #30) was isolated from extracts of New Zealand oysters (Tiostrea chilensis) [24] (Figure 1). Isolation of the active substance made it possible to determine its molecular formula as C 32 H 45 NO 4 (Table 1), whereas its absolute stereochemistry was later determined by X-ray crystallography (Figure 9) [99]. The origin of this toxin was first attributed to the dinoflagellate K. selliformis for two reasons: efflorescence of the latter was observed at the same time; gymnodimine was isolated in cultures of this species [100]. Two other analogs are known to be produced by K. selliformis: GYM-B (ID #31) and GYM-C (ID #32); whereas several others (12-methyl-GYM-A, 12-methyl-GYM-B, GYM-D, 16-desmethyl-GYM-D, GYM-E and more than 30 related gymnodimine-like compounds) have only been identified from Alexandrium ostenfeldii or Alexandrium peruvianum [59,87,[101][102][103][104]. The structure of GYM-B is similar to GYM-A, but contains an exocyclic methylene at position C17 and an allylic hydroxyl group at position C18 ( Figure 9). GYM-C is an oxidized isomer of GYM-B at position C-18. In 2015, it was demonstrated that GYM-A, GYM-B, and GYM-C could also be produced by A. ostenfeldii [105] ( Table 1). The use of phytoplankton net sampling allowed for the identification of GYM-A in Australia (Figure 1), together with K. selliformis cells (Moreton Bay, Queensland), and in China (Daya Bay) [45,46]. In terms of toxicity, GYM-A demonstrated high intraperitoneal toxicity in mice (LD 50 of 80-96 µg kg −1 ), while very low toxicity was reported after oral exposure [106,107]. Moreover, human toxicity is debatable, since no human poisoning could be associated with this toxin.

Biotransformation of Metabolites Produced by Karenia spp. in Shellfish
This section will only focus on the metabolism of BTXs and GYMs in shellfish. For both, individual analogs have been reported worldwide (Figure 1

Fatty Acid Conjugates of Brevetoxins
BTX-B3 is not a single compound, but a mixture of two fatty acid conjugates (myristic and palmitic acids) of the open D-ring of BTX-2 with oxidation of the terminal aldehyde (ID #60 and #61) ( Figure 11) [36]. Therefore, the use of the term BTX-B3 is discouraged from a chemical perspective (Table 2). These analogs were isolated from greenshell mussels (Perna canaliculus) involved in the NSP events of 1992-1993. Interestingly, the mixture of these fatty acid conjugates did not kill mice by intraperitoneal injection at a dose of 300 µg/kg [36]. No further fatty acid conjugates of BTXs (without amino-acid conjugation) were reported.

Fatty Acid Conjugates of Brevetoxins
BTX-B3 is not a single compound, but a mixture of two fatty acid conjugates (myristic and palmitic acids) of the open D-ring of BTX-2 with oxidation of the terminal aldehyde (ID #60 and #61) ( Figure 11) [36]. Therefore, the use of the term BTX-B3 is discouraged from a chemical perspective (Table 2). These analogs were isolated from greenshell mussels (Perna canaliculus) involved in the NSP events of 1992-1993. Interestingly, the mixture of these fatty acid conjugates did not kill mice by intraperitoneal injection at a dose of 300 µg/kg [36].  Figure 11. Chemical structures of fatty acid conjugates of brevetoxins (BTXs).

Fatty Acid Derivatives of Brevetoxin Amino acid Conjugates
In most cases, fatty acids react with amino acid-BTX conjugates (or peptide-BTX conjugates) through amide linkage to form a series of 19 fatty acid-amino acid-BTX conjugates (ID #58 and #70 to #87 presented in Figure 13). A greater number of amino acid-fatty acid conjugates of BTX-B have been identified compared to BTX-A. BTX-B4, the first isolated, is a mixture of N-tetradecanoyl (N-myristoyl) and N-hexadecanoyl (N-palmitoyl) conjugates with the cysteine sulfoxide moiety of BTX-B2 (ID #70 and #71, respectively) [36,40]. Therefore, as for BTX-B3 (ID #60 and #61), the use of the term BTX-B4 is discouraged. Based on MS/MS fragmentations, Wang et al. [39] reported numerous new analogs, all mentioned in Table 2. Several of these metabolites were confirmed in hard clams (Mercenaria sp.) naturally exposed to K. brevis blooms [22].

Other Brevetoxin Metabolites
Two metabolites, with m/z of 976.5 and 1004.5 (ID #59 and #88, respectively), have been identified from Eastern oysters (Crassostrea virginica) collected in the Gulf of Mexico [30,39]. These compounds have not been named. LC-MS/MS analysis revealed the presence of characteristic fragments of A-type and B-type backbone structure of BTXs, allowing Wang et al. [39] to propose structures and fragmentation pathways ( Figure 14). In both cases, the side chain possesses sulfoxide and carboxylic acid functions. These metabolites were much more abundant in field-exposed oysters compared with laboratory-exposed oysters. Since 2004, these BTX metabolites have never been reported in the literature.

Biomarkers of Brevetoxin Exposure in Shellfish
The great diversity of physicochemical properties of BTXs complicates the development of efficient multi-toxin analytical methods. Importantly, the products of shellfish metabolism have different lipophilic properties compared to their precursor. The binding of an amino acid or a peptide to the algal toxin contributes to decreased lipophilicity, while binding of a fatty acid increases it. Therefore, the extraction of metabolites with highly variable polarities constitutes a major difficulty. For this reason, different authors emphasized the relevance of use several metabolites as biomarkers of BTX exposure to monitor the toxicity of shellfish following a K. brevis bloom [4,18,23]. BTX-3 (ID #9), BTX-B5 (ID #13), taurine-BTX-B (BTX-B1; ID #62), cysteine-BTX-B (S-deoxy-BTX-B2, ID #63), and cysteine-BTX-B sulfoxide (BTX-B2, ID #64) constitute relevant candidates due to their relative persistence. Nevertheless, considering the toxin profile differences between shellfish species, the appropriate biomarkers should be chosen carefully depending on the monitored shellfish species. This high specificity may also complicate the broad application of biomarkers of exposure when different species are targeted. Recently, a strategy was successfully applied to monitor BTX-B in gastropods (Triplofusus giganteus, Sinistrofulgur sinistrum, Cinctura hunteria, Strombus alatus, Fulguropsis spirata) and in one species of clams (Macrocallista nimbosa) [17].  been identified from Eastern oysters (Crassostrea virginica) collected in the Gulf of Mexico [30,39]. These compounds have not been named. LC-MS/MS analysis revealed the presence of characteristic fragments of A-type and B-type backbone structure of BTXs, allowing Wang et al. [39] to propose structures and fragmentation pathways ( Figure 14). In both cases, the side chain possesses sulfoxide and carboxylic acid functions. These metabolites were much more abundant in field-exposed oysters compared with laboratory-exposed oysters. Since 2004, these BTX metabolites have never been reported in the literature.

Biomarkers of Brevetoxin Exposure in Shellfish
The great diversity of physicochemical properties of BTXs complicates the development of efficient multi-toxin analytical methods. Importantly, the products of shellfish metabolism have different lipophilic properties compared to their precursor. The binding of an amino acid or a peptide to the algal toxin contributes to decreased lipophilicity, while binding of a fatty acid increases it. Therefore, the extraction of metabolites with highly variable polarities constitutes a major difficulty. For this reason, different authors emphasized the relevance of use several metabolites as biomarkers of BTX exposure to monitor the toxicity of shellfish following a K. brevis bloom [4,18,23]. BTX-3 (ID #9), BTX-B5 (ID #13), taurine-BTX-B (BTX-B1; ID #62), cysteine-BTX-B (S-deoxy-BTX-B2, ID #63), and cysteine-BTX-B sulfoxide (BTX-B2, ID #64) constitute relevant candidates due to their relative persistence. Nevertheless, considering the toxin profile differences between shellfish species, the appropriate biomarkers should be chosen carefully depending on the monitored shellfish species. This high specificity may also complicate the broad application of biomarkers of exposure when different species are targeted. Recently, a strategy was successfully applied to monitor BTX-B in gastropods (Triplofusus giganteus, Sinistrofulgur sinistrum, Cinctura hunteria, Strombus alatus, Fulguropsis spirata) and in one species of clams (Macrocallista nimbosa) [17].

Monitoring Programs for Brevetoxins in Shellfish
In the United States, brevetoxins widely impact the coasts, especially in the Gulf of Mexico. The closure of production areas is based on Karenia spp. cell count or toxin analysis in shellfish tissues. A maximum limit of 20 mouse units per 100 g of tissue is applied using the mouse bioassay (MBA) and, with certain limitations, the enzyme linked immunosorbent assay (ELISA), as described in the National Shellfish Sanitation Program [114]. The re-opening of production areas is based on toxin analysis. With these analytical tools, the identity of the analogs remains unknown. In Australia, production area closures have different strategies depending on the federal states. MBA is implemented in Victoria and New South Wales; however, an LC-MS/MS analysis is also carried out in New South Wales [115,116]. BTX-2 (ID #8), BTX-3 (ID #9), BTX-B5 (ID #13), taurine-BTX-B (BTX-B1; ID #62), cysteine-BTX-B (S-deoxy-BTX-B2; ID #63), and cysteine-BTX-B sulfoxide (BTX-B2; ID #64) are the BTX analogs sought [117]. In New Zealand, BTXs are regulated (maximum permissible level of 0.8 mg BTX-2 equivalent per kg of shellfish meat) and the reference method is based on LC-MS [118]. In France, BTX-2 and BTX-3 have been monitored since 2018 [10] as part of a program on emerging toxins in marine shellfish (EMERGTOX). Recently, a working group was set up by the French Agency for Food, Environmental and Occupational Health and Safety (ANSES) to prevent health risks associated with the consumption of shellfish contaminated with BTXs [11,119]. A guidance level of 180 µg BTX-3 eq./kg shellfish meat was established, considering a protective default portion size of 400 g shellfish meat. A monitoring strategy has also been proposed and is divided into three parts [119]. First, the implementation of an ELISA test for the screening of B-type brevetoxin metabolites in shellfish flesh. This approach considers brevetoxins not quantifiable by LC-MS/MS due to the lack of commercial standards. Second, the targeted physicochemical analysis of brevetoxins for which standards are available with an LC-MS/MS method developed for a wide range of polarity. Based on available reference materials and toxicity data, a list of analogs to seek first was established and includes:

Gymnodimines (GYMs)
Several studies have demonstrated the accumulation of gymnodimines (GYMs) in different shellfish species (Table 3). The discovery of GYM-A (ID #30) from New Zealand oysters (Tiostrea chilensis) in 1995 immediately highlighted this capacity [24]. A few years later, several batches of clams harvested in Tunisia were neurotoxic to mice by i.p. [27]. Analytical investigations allowed for the unequivocal identification of GYM-A. Several studies on Tunisian samples followed [25,26]. In 2012, GYM-A, GYM-B (ID #31) and GYM-C (ID #32) were detected in the same shellfish species collected in the Gulf of Gabes, along with the presence of K. selliformis [25]. Other studies were carried out in South Africa, Australia, and China ( Figure 1) and allowed for the detection of GYM-A in oysters, gastropods, mussels, donax and pen shells [43][44][45][46][47]. In 2013, numerous gymnodimine fatty acid conjugates were identified from grooved carpet shells (Ruditapes decussatus) collected in Tunisia (Gulf of Gabes) [42]. In total, 46 fatty acid conjugates of GYM-A (ID #89 to ID #134) have been reported, with acyl carbon chain lengths comprised between C12:0 and C24:6 ( Table 3). The presence of a series of GYM-B and GYM-C fatty acid conjugates was also established, with esters of 16, 18, 20 and 22 carbon chain lengths being the most abundant. However, only the 18:0 esters of GYM-B and GYM-C were specifically identified, with respective ID number of #135 and # ID136 ( Figure 15). Some of the GYM-A fatty acids conjugates were also identified in shellfish collected in the South China Sea [41]. Clam (Antigona lamellaris) and pen shell (Atrina pectinata) profiles were different. In clams, the 18:0 ester was the most abundant, followed by the 20:1 ester, whereas the 20:1, 18:0 and 22:2 esters dominated the acyl ester profile in pen shell samples. These authors also experimentally exposed mussels (Mytilus galloprovincialis) to K. selliformis for 96 h. In all, 28 fatty acids esters were detected with acyl carbon chain lengths ranging between C14:0 and C24:6. Interestingly, GYM-A was detected in oysters (Crassostrea sp.) and gastropods (Batillaria zonalis), but no GYM-A esters were observed in these species. Further investigations are required to obtain better knowledge of GYM toxicity. However, considering the significant presence of GYM acyl esters, more occurrence data are required to conduct a risk assessment at a later stage.     [42] fatty acids esters were detected with acyl carbon chain lengths ranging between C14:0 and C24:6. Interestingly, GYM-A was detected in oysters (Crassostrea sp.) and gastropods (Batillaria zonalis), but no GYM-A esters were observed in these species. Further investigations are required to obtain better knowledge of GYM toxicity. However, considering the significant presence of GYM acyl esters, more occurrence data are required to conduct a risk assessment at a later stage.

Accumulation and Biotransformation of Metabolites Produced by Karenia spp. in Other Marine Organisms
Blooms of K. brevis are commonly associated with massive mortality for marine organisms. Massive death episodes of fish (prey fishes, sharks, rays), mammals (dolphins, manatees) or shorebirds, all occurring in Florida between 1999 and 2006, were associated with K. brevis blooms and the presence of BTXs [5,49,50,120]. Data regarding the accumulation and biotransformation of the other groups of toxins produced by Karenia spp. are lacking. In several studies, the implementation of biochemical methods of analysis, such as enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), or receptorbinding assay (RBA), made it possible to measure high levels of BTX-like toxins in animal tissues, but did not enable identification of the specific nature of the BTX analogs. In a few studies conducted on samples collected in the Gulf of Mexico (Figure 1), LC-MS and LC-MS/MS analyses were implemented to obtain this information [2,[48][49][50].
After major dolphin (Tursiops truncatus) mortality events between 1999 and 2006 in Florida, BTXs and domoic acid (amnesic toxin produced by the Pseudo-nitzschia diatom) were sought in dead animals [50]. Of the 105 animals collected in 2004, 100% of the tested animals were positive for BTXs and 89% for domoic acid. Moreover, dolphin stomach contents frequently consisted of BTX-contaminated menhaden (Brevoortia sp.). For the period 2005-2006, 93% of the 90 dolphins were positive for BTXs, whereas domoic acid was not detected in these animals. Among the samples analyzed by LC-MS, BTX-3 (ID #9) was the predominant toxin in the four stomach contents tested (fishes partially digested) and was also detected in 10 of the 12 livers tested. Two peaks co-eluting with BTX-2 were not attributed to any known toxin.
An experimental study on shrimps (Litopenaeus vannamei) demonstrated that these marine organisms could represent a vector of BTXs [112]. The authors found by LC-MS levels of 80 and 90 µg eq BTX-2/kg in the digestive glands and muscles, respectively. It should be noted that this contamination started to appear in the digestive gland and the muscle after 30 and 45 days of exposure, respectively, and with relatively low cellular concentrations of K. brevis (10 3 to 10 6 cells/L). Macrobenthic invertebrates (including polychaetes and crustaceans belonging to amphipods and isopods) collected during a K. brevis bloom in Florida also demonstrated their capacity to accumulate BTXs [2]. However, the identity of the analogs was not determined due to the analytical tool implemented for the analysis (ELISA test).

Conclusions
In this review, we summarized and gathered the knowledge reported in the literature over recent decades of over 130 potentially toxic metabolites reported from Karenia spp. or metabolized by marine organisms. The structures and certain properties of these metabolites are presented. This information could constitute an interesting basis in view of library-building, particularly for physicochemical analysis using liquid chromatography combined with high-resolution mass spectrometry (LC-HRMS). Such an approach could be an interesting tool to simultaneously screen the different analogs mentioned here, and could also allow us to discover new analogs. The accumulation and biotransformation by marine organisms of the metabolites produced by Karenia spp. was also assessed in depth for BTXs and GYMs. We point out the need for more information on BTXs in marine organisms other than shellfish. Occurrence data would also be valuable for several groups of potentially toxic metabolites discovered in the last two decades. These data, combined with additional toxicology assays, could be used to carry out a risk assessment. Lastly, we compiled all the names reported for metabolites in the existing literature. Among the different nomenclatures, we tried to select the terms that allowed us to obtain a better overall consistency, hoping that this work will be useful to harmonize practices in the future.